COGNAC: COnnection tables Graphs for Nucleic ACids

The COGNAC server is a graph theoretical program that analyses and annotates hydrogen bonded interactions between bases in RNA 3D structures (PDB formatted). A major component of RNA structure stabilization are the hydrogen bonded interactions between the base residues. The importance and biological relevance for large clusters of base interactions can be much more easily investigated when their occurrences can be systematically detected, catalogued, and compared.

The COGNAC server first runs a hydrogen bond calculation program to determine the hydrogen bonds. Currently the user also has the option of using either a database containing both protonated and non-protonated base residue reference libraries. The connectivity of the hydrogen bonds are then transferred into a connection table. COGNAC, which is a graph theoretical implementation of the Ullmann subgraph isomorphism, then solves the problem of matching a query connection table graph representation to its matching subgraph; here, the search queries represent possible base interactions and the relationships are mapped by a graph theoretical algorithm.


See example of a connection table for a base triple here: PDF | PNG image.

Precomputed example:
A search of all patterns in 1FFK, the large ribosomal subunit from Haloarcula marismortui. [See example]

Demonstration run (1 structure):
Click HERE to load a prefilled display of a search for triples in 1FFK. The display will appear in the frame below (please scroll down). Click "submit" to see the result.

Demonstration run (2 structures):
Click HERE to load a prefilled display of a search for all patterns in 4JUX and 4BTD. The display will appear in the frame below (please scroll down). Click "submit" to see the result.

Citation for referencing COGNAC:
1. Mohd Firdaus-Raih, Hazrina Yusof Hamdani, Nurul Nadzirin, Effirul Ikhwan Ramlan, Peter Willett and Peter J. Artymiuk. 2014. COGNAC: a web server for searching and annotating hydrogen-bonded base interactions in RNA three-dimensional structures. Nucleic Acids Res. 42 (W1): W382-W388. HTML, PDF

To run a search:
STEP 1. Select to search either the option of (1) searching for a specific set of hydrogen bonded base connections in a database of RNA structures (sourced from the PDB) or (2) to upload an RNA structure (in PDB format) to annotate the hydrogen bonded base interactions present in the query structure.
STEP 2. Select the number of residues and the arrangement of the interactions for the query patterns from the list provided (2-6 bases or select "All patterns" for arrangement types).
STEP 3. For specific patterns selected in STEP 2; enter the specific residue or general type (for all base possibilities).

For searches using Option 3; the difference is mainly in the requirement to have two structures uploaded. Please note that the program does not filter out highly dissimilar structures which are not useful comparisons.

Search times and output visualization:
1. Each search may take time to complete depending on the complexity of the query and the server load (usually completed within 2 minutes). Users are advised to bookmark the results page or copy the link to the results for analysis at a later point in time if a search takes more than 10 minutes.
2. Please note that COGNAC output can be visualized using JSMol. This bypasses issues with Java such as the block by the Mac OSX operating system when using Jmol.
3. Search results will be kept for at least 2 days. The main search results can be downloaded to your local computer while still retaining the server side ability to view the hits in Jmol.

Please email for help or if you've detected any problems with the search.

Computational resources provided by the Genome Computing Centre, Malaysia Genome Institute